I spend a lot of time these days talking with companies about the need for a formal approach to enabling what is often called “discovery analytics” or “exploratory analytics.” What I find is that many people have a fundamental misunderstanding of what discovery analytics is all about. There is one analogy that I have found to be effective in getting people to better understand the concept. In this blog, I’ll walk you through that analogy.
I spend a lot of time these days talking with companies about the need for a formal approach to enabling what is often called “discovery analytics” or “exploratory analytics.” What I find is that many people have a fundamental misunderstanding of what discovery analytics is all about. There is one analogy that I have found to be effective in getting people to better understand the concept. In this blog, I’ll walk you through that analogy.
IT ISN’T AIMLESS HACKING!
Many people get very concerned when I begin to discuss discovery analytics as being not fully defined, constantly evolving, and remaining fluid. They tell me that what I’m saying sounds a lot like a mad scientist sitting down running random experiments in the hope of finding something useful. I do not espouse such an approach, I can assure you!
On the contrary, a discovery process should always start with a specific high priority business problem in mind. There should also be at least a general idea of how to address the problem effectively through analytics after some initial brainstorming. At that point, a discovery process is started to explore how well our ideas actually do address the problem. Typically, a discovery process involves one or more big unknowns:
- We may be addressing a totally new business problem
- We may be utilizing one or more new and /or largely untested data sources
- We may be making use of analytics techniques that we haven’t used in the past
In short, while we think a proposed approach has merit, we really don’t know for sure how well it will work. The only way to find out is to dig in and see what we find. As a part of that process, we may well adjust our approach across any number of dimensions. The final solution we find may be somewhat different than the path we started down, but it will be found by remaining focused on the core problem we start with.
WE’RE REALLY TALKING ABOUT RESEARCH & DEVELOPMENT
If analytics is going to be a strategic component of your business, then you need to invest in analytics just like you do for other core products and services your company offers. Discovery analytics aren’t mindless hacking any more than traditional research and development activities are. People accept that an R&D team will have to be creative and try many paths to identify a winner. At first glance, some say that this is acceptable in a traditional product setting but not in a discovery analytics setting. I suggest that it is important to realize that the two are the same.
One of our high-tech customers takes only take a handful of new products to market in any given year. However, massive investments with many trial and error experiments happen behind the scenes to get those products ready. At the other end of the product complexity scale, quick-serve companies only bring a few new menu items to market each year. In their test kitchens, however, a never-ending stream of new recipes and ingredients are explored to arrive at the winners that make it to your local restaurant.
Discovery analytics are much the same. A variety of attempts will be made, of which only a few will turn out worthy of a full deployment. But, those that are worthy will have a high level of strategic value if efforts are focused in the correct areas of the business up front. Some ideas may not work out at all and will have to be abandoned. I’m sure there are also many computer chips or chicken sandwiches that never made it out of the R&D process either. Not every effort will turn into a winner, but working through the losers is the only way to get to the winners. In fact, if your analytics do not generate some losers, something is wrong: either they aren’t being made known, or the analytics team is not continuously pushing the limits of innovation.
RAMPING UP YOUR ANALYTICS R&D FUNCTION
It is critical to redirect discussion on discovery analytics toward the R&D parallel. People are much more comfortable with R&D because it is viewed as a rational, scientific, disciplined approach to developing new products and ideas. Discovery analytics is just that. Of course, it takes ongoing effort to ensure that any R&D function stays on the right path. With correct oversight and leadership, however, there is no reason your organization can’t reap large benefits from discovery analytics over time.
The best thing about building up an analytics R&D function is that it will self-propel itself forward. Start by getting commitment for a limited number of human and technology resources to address only a handful of critical business problems. Once you prove it works, slowly ask for more resources to attack more problems. Over time, you’ll be able to grow a stable, accepted analytics research and development function. You just have to push through the initial misunderstandings and the resistance to the unknown.
I think you’ll find, like I have, that very few people will argue against the merits of research and development as a business endeavor. The first step in the process of enabling discovery analytics at your organization is to ensure that people understand that you’re just talking about a different type of R&D effort. You’ll still have a lot of work to do, but at least people will be willing to listen to you make your case and hopefully give you a chance to show what you can do.